Commit a9d18c4a0c2be3d5e7dcedbedc32a9998b1e5515 | zig

author	Matthew Lugg <mlugg@mlugg.co.uk> 2026-02-18 14:51:59+01:00
committer	Alex Rønne Petersen <alex@alexrp.com> 2026-02-21 23:39:34+01:00
log	a9d18c4a0c2be3d5e7dcedbedc32a9998b1e5515
tree	481a178e0ddef02dffa6b74dc8d1d0cb5d0f7665
parent	c8dd05030519fc52dd457531f6b2654e1ad129b2
signature	Signed by SSH key SHA256:7B/LJ7bpR1eX8aCXSr4mtd5M45VMPKcx9zY8e95b5QM

std.heap.PageAllocator: avoid mremaps which may reserve potential stack space

Linux's approach to mapping the main thread's stack is quite odd: it essentially
tries to select an mmap address (assuming unhinted mmap calls) which do not
cover the region of virtual address space into which the stack *would* grow
(based on the stack rlimit), but it doesn't actually *prevent* those pages from
being mapped. It also doesn't try particularly hard: it's been observed that the
first (unhinted) mmap call in a simple application is usually put at an address
which is within a gigabyte or two of the stack, which is close enough to make
issues somewhat likely. In particular, if we get an address which is close-ish
to the stack, and then `mremap` it without the MAY_MOVE flag, we are *very*
likely to map pages in this "theoretical stack region". This is particularly a
problem on loongarch64, where the initial mmap address is empirically only
around 200 megabytes from the stack (whereas on most other 64-bit targets it's
closer to a gigabyte).

To work around this, we just need to avoid mremap in some cases. Unfortunately,
this system call isn't used too heavily by musl or glibc, so design issues like
this can and do exist without being caught. So, when `PageAllocator.resize` is
called, let's not try to `mremap` to grow the pages. We can still call `mremap`
in the `PageAllocator.remap` path, because in that case we can set the
`MAY_MOVE` flag, which empirically appears to make the Linux kernel avoid the
problematic "theoretical stack region".

1 files changed, 4 insertions(+), 1 deletions(-)

lib
- std
  - heap
    PageAllocator.zig

lib/std/heap/PageAllocator.zig+4-1

 @@ -225,7 +225,10 @@ pub fn realloc(uncasted_memory: []u8, alignment: Alignment, new_len: usize, may_
     if (new_size_aligned == page_aligned_len)
         return memory.ptr;
-    if (posix.MREMAP != void) {
+    // When the stack grows down, only use `mremap` if the allocation may move.
+    // Otherwise, we might grow the allocation and intrude on virtual address
+    // space which we want to keep available to the stack.
+    if (posix.MREMAP != void and (stack_direction == .up or may_move)) {
         // TODO: if the next_mmap_addr_hint is within the remapped range, update it
         const new_memory = posix.mremap(memory.ptr, page_aligned_len, new_size_aligned, .{ .MAYMOVE = may_move }, null) catch return null;
         return new_memory.ptr;